EMG-Controlled Soft Robotic Bicep Enhancement

Abstract

Industrial workers often engage in repetitive lifting tasks. This type of continual loading on their arms throughout the workday can lead to muscle or tendon injuries. A non-intrusive system designed to assist a worker's arms would help alleviate strain on their muscles, thereby preventing injury and minimizing productivity losses. The goal of this project is to develop a wearable soft robotic arm enhancement device that supports a worker's muscles by sharing the load during lifting tasks, thereby increasing their lifting capacity, reducing fatigue, and improving their endurance to help prevent injury. The device should be easy to use and wear, functioning in relative harmony with the user's own muscles. It should not restrict the user's range of motion or flexibility. The human arm consists of numerous muscles that work together to enable its movement. However, as a proof of concept, this project focuses on developing a prototype to enhance the biceps brachii muscle, the primary muscle involved in pulling movements during lifting. Key components of the prototype include a soft robotic muscle or actuator analogous to the biceps, a control system for the pneumatic muscle actuator, and a method for securing the soft muscle to the user's arm. The McKibben-inspired pneumatic muscle was chosen as the soft actuator for the prototype. A hybrid control algorithm, incorporating PID and model-based control methods, was developed. Electromyography (EMG) and pressure sensors were utilized as inputs for the control algorithms. This paper discusses the design strategies for the device and the preliminary results of the feasibility testing. Based on the results, a wearable EMG-controlled soft robotic arm augmentation could effectively enhance the endurance of industrial workers engaged in repetitive lifting tasks.

Keywords: EMG sensor; McKibben pneumatic muscle; PID control; soft robotics.

Figure 1

Pneumatic muscle: (a) working principle; and (b) four prototypes labeled #1 to #4.

Figure 2

Pneumatic muscle testing platform and setup.

Figure 3

Relationship between pressure, muscle contraction, and loads from muscle tests (Note: M1–5 lbs (2.27 kg) refers to muscle #1 with a 5 lbs (2.27 kg) load).

Figure 4

Muscle pressure control algorithm flowchart.

Figure 5

(a) Test datasets of EMG readings and various lift weights; and (b) the relationship between the weights lifted and maximum EMG readings (* indicates the maximum EMG reading from the test dataset at a specified weight).

Figure 6

EMG readings for 20 repetitions with a lift weight of 20 lbs.

Figure 7

(a) Test results on relationship between pressure ratio and duty cycle; and (b) quadratic fit of pressure ratio–duty cycle relationship overlaid on the test data.

Figure 8

Effect of muscle length on the smoothness of the pressure ratio–duty cycle curve.

Figure 9

(a) Pressure ratio vs. duty cycle curves for various PWM frequencies (two red vertical lines indicate the duty cycle range at 100 Hz PWM frequency and two black vertical lines indicate the duty cycle range at 25 Hz PWM frequency); and (b) mean squared error of quadratic fit at various PWM frequencies (* indicates mean squared error at specified PWM frequency (Hz)).

Figure 10

Control system electronic wiring diagram.

Figure 11

Pneumatic control system diagram.

Figure 12

Wearable device prototype testing setup illustration: (a) front; and (b) back.

Figure 13

(a) User wearing physical prototype for testing; and (b) industrial air supply to the prototype.

Figure 14

Control box with electronic components and controls.